gen has been shown to naturally occur at low levels in nor-mal adult skin, during embryonic development, and duringwound healing.9-11 Homotrimeric type I collagen has alsobeen associated with certain forms of Ehlers-Danlos syn-drome (EDS)12-14 and osteogenesis imperfecta (OI).15,16

Altered elastin content and structure have been impli-cated in the formation of abdominal aortic aneurysms.2,17

Because the tensile strength of collagen (1.0 × 109 N/m2)is approximately 1000 times greater than that of elastin(4.6 × 105 N/m2), defects in collagen should alter vascu-lar integrity and contribute to aneurysmal development.18

Defects in type III collagen, responsible for EDS type IV,have a strong association with aortic aneurysmal degener-ation.19-21 However, the role of type I collagen in aorticwall integrity and aneurysmal development remains to beelucidated. This study investigated the impact of theabsence of the α2(I) collagen chain and/or the presenceof homotrimeric type I collagen [α1(I)]3 on the biome-chanical properties of the thoracic aorta to begin to eluci-date the causative role and/or contributory role of type Icollagen defects and/or polymorphisms in the develop-ment of vascular disease.

MATERIAL AND METHODS

Animals. Mice who were homozygous for thehomotrimeric type I collagen (oim/oim), heterozygousmice (oim/+), and wildtype mice (+/+) were purchasedfrom Jackson Laboratory (Bar Harbor, Me). Additional

Elastin and collagen are the primary extracellularmatrix components of the aortic wall, providing theintegrity to withstand the outward forces exerted by arte-rial pressure.1,2 These elements make up approximately50% of the dry weight of normal arteries.1 The dominantcomponent in the thoracic aorta is elastin (approximately60%), whereas the composition is reversed in the abdomi-nal aorta (70% collagen).1 Arterial collagen is composedpredominantly of type I and type III.3

The normal type I collagen isotype is a heterotrimericmolecule, composed of two α1(I) chains and one similar,but genetically distinct, α2(I) chain.4,5 The homotrimerictype I collagen isotype, composed of three α1(I) chainswas originally discovered in certain tumors and culturedcancer cell lines.6-8 Since then, homotrimeric type I colla-

1263

From the Division of Vascular Surgery,a and the Departments ofBiochemistry,b Biological and Agricultural Engineering,c and ChildHealth,d University of Missouri, Columbia.

Competition of interest: nil.A portion of this study was presented at the Seventh International

Conference on Osteogenesis Imperfecta, Montreal, Canada, Aug29–Sep 1, 1999.

Reprint requests: Charlotte L. Phillips, PhD, Departments of Biochemistryand Child Health, Division of Medical Genetics, University of Missouri-Columbia, M121 Medical Sciences Bldg, Columbia, MO 65212 (e-mail: PhillipsCL@missouri.edu).

Copyright © 2001 by The Society for Vascular Surgery and The AmericanAssociation for Vascular Surgery.

0741-5214/2001/$35.00 + 0 24/1/113579doi:10.1067/mva.2001.113579

The role of type I collagen in aortic wallstrength with a homotrimeric [α1(I)]3collagen mouse modelAngela G. Vouyouka, MD,a Brent J. Pfeiffer, BS,b Timothy K. Liem, MD,a Timothy A. Taylor, PhD,cJunaid Mudaliar, MD,a and Charlotte L. Phillips, PhD,b,d Columbia, Mo

Purpose: Elastin and collagen (types I and III) are the primary load-bearing elements in aortic tissue. Deficiencies andderangements in elastin and type III collagen have been associated with the development of aneurysmal disease.However, the role of type I collagen is less well defined. The purpose of this study was to define the role of type I col-lagen in maintaining biomechanical integrity in the thoracic aorta, with a mouse model that produces homotrimerictype I collagen [α1(I)]3, rather than the normally present heterotrimeric [α1(I)]2 α2(I) type I collagen isotype.Methods: Ascending and descending thoracic aortas from homozygous (oim/oim), heterozygous (oim/+), and wildtype(+/+) mice were harvested. Circumferential and longitudinal load-extension curves were used as a means of determin-ing maximum breaking strength (Fmax) and incremental elastic modulus (IEM). Histologic analyses and hydroxypro-line assays were performed as a means of determining collagen organization and content.Results: Circumferentially, the ascending and descending aortas of oim/oim mice demonstrated significantly reducedFmax, with an Fmax of only 60% and 23%, respectively, of wildtype mice aortas. Oim/oim descending aortas demon-strated significantly greater compliance (decreased IEM), and the ascending aortas also exhibited a trend towardincreased compliance. Reduced breaking strength was also demonstrated with longitudinal extension of the descendingaorta.Conclusion: The presence of homotrimeric type I collagen isotype (absence of α2(I) collagen) significantly weakens theaorta. This study demonstrates the integral role of type I collagen in the biomechanical and functional properties ofthe aorta and may help to elucidate the role of collagen in the development of aneurysmal aortic disease or dissection.(J Vasc Surg 2001;33:1263-70.)

+/+ mice (B6C3Fe a/a) were purchased to serve as addi-tional control animals, because the oim mutation is main-tained in the B6C3Fe a/a background.22 All animals werehoused in a facility approved by the American Associationfor Accreditation of Laboratory Animal Care (AAALAC),provided free access to water and food (standard labora-tory diet, autoclavable rodent laboratory chow, 5010,Purina Mills, Richmond, Ind), and handled according toan approved protocol and the approved regulations of thefacility. Oim/oim, oim/+, and +/+ genotypes were con-firmed by means of polymerase chain reaction-restrictionfragment length polymorphism analysis.23 Mice (age, 4-7months) were euthanized with CO2 asphyxiation for bio-mechanical and histologic analyses.

Preparation of the aortic segments. Aortas were har-vested with a Weck operating microscope (J. K. HopplCorp, Lindenhurst, NY) with a lens magnification of 15×.The chest of the animals was opened via a median ster-notomy, and the thoracic aorta was dissected from the sur-rounding adipose tissue and harvested from the aortic rootto the diaphragmatic crurae. For circumferential analyses,two 5-mm long aortic rings were prepared: a proximal ringincluding aortic root and aortic arch and a distal aortic ringincluding the descending aorta just below the take off of theleft subclavian artery. A pair of 0.5-mm diameter stainlesssteel hooks was inserted in each aortic ring. For longitudinalanalyses, the whole segment of the descending aorta, fromthe left subclavian artery to the diaphragm, was excised foruse in the longitudinal extension analyses. The infrarenalaorta was not analyzed because of its decreased size (< 5 mmin length and < 1 mm in diameter).

Biomechanical studies. To obtain circumferentialload-extension curves, we attached each aortic ring to aTA-XT2 texture analyzer (tensile testing apparatus;Texture Technologies, Scarsdale, NY), placed it in a Krebssolution bath at room temperature (24 ± 1°C), andstretched it at a constant tensile speed of 0.1 mm/s (Fig1). During the aortic ring extension (length [l, in mil-limeters]), the required tensile load (Force [F, in grams])was recorded at 0.01-second intervals. The maximal ten-sile load (Fmax) is the greatest load (force) before vesselbreakage (Fig 1, value indicated by C). While the aorticsegment is deformed by tensile load, the wall exerts anincreasing retractive force, aiming to a final equilibrium.Stiffness of the aortic ring results from the relationshipbetween deformation (extension) and tensile load and canbe expressed as:

E = f/l

in which E is the elastic modulus, f is the tensile load, andl is the extension. Because of the nonlinear relationship,the incremental elastic modulus (IEM = ∆f/∆l) was usedas a means of comparing the stiffness of the aortic seg-ments among the genotypes. The IEM was calculated tobe the ratio of the incremental load (in grams) to theincremental extension (in millimeters). The IEM was esti-mated by performing a linear regression analysis on eachload-extension curve. Best-fit lines were obtained at lowand high (1.5× to 2× diameter) degrees of extension in theaortic rings. Elastin bears most of the load at low degreesof extension, whereas collagen bears most of the load athigher degrees. The slope of the regression line at highdegrees of extension was used to estimate the IEM,because this is the region in which differences in collagenstructure are observed. IEM and Fmax values were stan-dardized to dry mass weight of the respective aortic sam-ple (in milligrams).

In the longitudinal direction, aortic segments of thewhole descending aorta were extended at 0.1 mm/s alongthe longitudinal axis (axial stretching), and the Fmax and theIEM were determined. Fig 1 represents schematically theTA-XT2 texture analyzer used for the aortic tissue extension.

Hydroxyproline assay. The hydroxyproline contentwas used as a means of determining the total amount ofcollagen in the aortic wall. Hydroxyproline represents 20%to 30% of the amino acids of mature collagen molecules.4,5

After load-extension curve analyses, the aortic tissue wasdried and weighed, and the total hydroxyproline contentwas determined by means of a colorimetric assay (on thebasis of the oxidation of hydroxyproline to a compoundthat reacts with p-dimethylaminobenzaldehyde to form achromophore) established by Stegemann and Stadler.24

Histologic analysis. Aortic sections were harvestedfor histologic and morphologic analyses from 2 oim/oimand 3 +/+ mice. Aortas were perfused at 112 mm Hg, apressure that mimics the normal oim/oim mouse systolicpressure as determined by means of tail-cuff plethysmog-raphy. Perfusion was performed through the left ventricle

JOURNAL OF VASCULAR SURGERY1264 Vouyouka et al June 2001

Fig 1. Schematic representation of TA-XT2 texture analyzer sys-tem used for load-extension studies of the aortic samples. TheTA-XT2 is a computer-controlled system that regulates the speedof extension (0.1 mm/s) and measures the loading force (ingrams) for each extension point (in millimeters). For circumfer-ential extension analyses, a pair of stainless steel 0.5-mm hookswere threaded into aortic rings and secured to the TA 96 systemclamps. For longitudinal extension analyses, aortic segments weresecured in the TA 96 system clamps directly.

with 0.9% NaCl (2 minutes), followed by neutral buffered10% formalin fixation (3 minutes). Aortas were then har-vested and processed for histologic examination.Transverse sections were collected from the aortic arch,proximal descending thoracic aorta, and the proximalabdominal aorta. Five-micron sections of each aorta werestained with hematoxylin and eosin as a means of qualita-tively assessing cell number and morphology, picrosiriusred as a means of assessing collagen, and Verhoeff-vanGieson as a means of assessing elastin.

Statistical analysis. Differences in Fmax and IEMbetween oim/oim, oim/+, and +/+ mice were evaluated bymeans of analysis of variance and analysis of covariance,with age as the covariant. Statistical significance wasdefined as a P value less than .05.

RESULTS

Six oim/oim, 8 oim/+, and 6 +/+ mice were sacrificedfor the circumferential analyses. The mean body weightsof the +/+, oim/+, and oim/oim mice were 25.0 ± 2.1 g,26.3 ± 2.6 g, and 20.8 ± 2.1 g, respectively. Althoughoim/oim mice weighed 21% and 17% less than the oim/+(P = .001) and +/+ (P = .02) mice, there appeared to beno differences in aortic diameters or lengths.

Biomechanical studies. Representative circumferen-tial load extension curves for the descending aortic ringsare depicted in Fig 2. Complete disruption of vessels wasnoted at ring extensions that were 1.5 to 2 times the ini-tial diameters for all genotypes (indicated by the Fmaxvalue of C). As demonstrated by means of Fig 2, thecurves were nonlinear. At low degrees of stretch, smallload increments resulted in large-size deformation (seg-ment AB in the curve). At high degrees of ring extension(1.3 to 2 times the diameter of the ring) in which collagen

JOURNAL OF VASCULAR SURGERYVolume 33, Number 6 Vouyouka et al 1265

bears most of the load, the same load increments resultedin much smaller deformation (segment BC). Linearregression analysis of segment BC lead to the best-fit line(inset of Fig 2, a, b, and c), which was used as a means ofestimating the IEM.

Fig 3 summarizes the results from the load-extensioncurves obtained from circumferential extension of theascending aortas. Circumferentially, the maximum ten-sile load (Fmax) of the oim/oim ascending aorta (132 ±42 g/mg) was significantly reduced, as compared withthe Fmax of the +/+ ascending aorta (219 ± 75 g/mg, P = .04). The oim/+ mice demonstrated intermediatevalues for Fmax (185 ± 66 g/mg), although the resultswere not significantly different from the oim/oim or +/+groups (Fig 3, a). When the stiffness of the ascendingaorta was evaluated, the IEM values of oim/oim (IEM =379 ± 332 g/mm/mg), oim/+ (IEM = 309 ± 126g/mm/mg), and +/+ (IEM = 417 ± 244 g/mm/mg)mice were not statistically different. However, there wasa trend for the oim/oim and oim/+ aortas to be moreextensible (lower IEM values) than the +/+ ascendingaortas (Fig 3, b).

In contrast, circumferential extension of the descendingaortas demonstrated significant differences in both thebreaking strength and stiffness between oim/oim, oim/+,and +/+ mice (Fig 4). The maximal load (Fmax) required tocircumferentially disrupt the descending aortic ring of +/+mice was 447 ± 124 g/mg, whereas the Fmax of oim/+ andoim/oim mice descending thoracic aortas were 242 ± 112g/mg (P = .002) and 103 ± 42 g/mg (P = .001), respec-tively (Fig 4, a). Similarly, the +/+ mice had stiffer (lessextensible) distal aortic rings (IEM, 788 ± 128 g/mm/mg)than the oim/+ (486 ± 271 g/mm/mg, P = .01) andoim/oim (188 ± 70 g/mm/mg, P = .0001) mice (Fig 4, b).

Fig 2. Representative circumferential load–extension curves obtained from the descending aortic rings of wildtype (+/+; a), heterozy-gous (oim/+; b), and homozygous (oim/oim; c) mice. The AB segment represents low-degree extension, and the BC segment representshigh-degree extension (1.5-2 × initial aortic diameter). Inset is linear analysis and best-fit line of segment BC. Incremental elastic mod-ulus was calculated from slope of BC segment, IEM = df/dl.

a b c

JOURNAL OF VASCULAR SURGERY1266 Vouyouka et al June 2001

Fig 3. Analyses of circumferential maximum breaking strength (Fmax; a) and incremental elastic modulus(IEM; b) of the ascending aorta of wildtype (+/+; n = 6; black bar), heterozygous (oim/+; n = 8; diago-nal bar), and homozygous (oim/oim; n = 6; open bar) mice. Values are expressed as the mean plus or minusSD in grams (Fmax) and grams per millimeter (IEM) per milligram of dried aortic tissue. *P ≤ .05.

Fig 4. Analyses of circumferential maximum breaking strength (Fmax; a) and incremental elastic modulus(IEM; b) of descending aorta of wildtype (+/+; n = 6; black bar), heterozygous (oim/+; n = 8; diagonalbar), and homozygous (oim/oim; n = 6; open bar) mice. Values are expressed as mean ± SD in grams (Fmax)and grams per millimeter (IEM) per milligram of dried aortic tissue. *P ≤ .01 ; **P ≤ .001.

a b

a b

Longitudinal extension studies of the descending tho-racic aortas were performed in 5 oim/oim, 3 oim/+, and 7+/+ mice. The Fmax of the +/+ mice (41.7 ± 13.5g/mm/mg) was significantly greater than that of theoim/oim (16.4 ± 8.5 g/mg, P = .003) and oim/+ mice(13.11 ± 5.2 g/mg, P = .017; Fig 5, a). However, no sig-nificant difference was noted in the longitudinal extensi-bility (IEM) of the +/+ descending aorta (10.7 ± 1.4g/mm/mg), as compared with the oim/oim descendingaorta (9.8 ± 3.3 g/mm/mg, P = .512; Fig 5, b). The IEMof the oim/+ mice (5.9 ± 1.6 g/mm/mg) did appear dif-ferent from that of the +/+ mice (P = .034), but the realsignificance of this is unclear and may reflect, in part, thesmall sample number of oim/+ mice (n = 3).

Hydroxyproline assay. Hydroxyproline content (ameasure of total collagen) was measured in 6 +/+, 3oim/+, and 4 oim/oim mice. There were no significant dif-ferences in the amount of total collagen present in thedescending thoracic aortas of the +/+ (13.5 ± 3.8 µg collagen/mg dry tissue), oim/+ (12.0 ± 1.6 µg/mg), andoim/oim (12.3 ± 3.0 µg/mg) mice. This suggests that the altered biomechanical properties may not simplyreflect quantitative differences in total collagen, but morequalitative alterations in the fibrillar organization andarchitecture.

Histologic analysis. Preliminary histologic evalua-tion of the descending thoracic aorta was performed in

JOURNAL OF VASCULAR SURGERYVolume 33, Number 6 Vouyouka et al 1267

2 oim/oim and 3 +/+ mice (Fig 6). Although only a lim-ited number of animals were examined histologically,examination of hematoxylin and eosin–stained sectionssuggested there were no lesions or major differences in cellnumbers. Furthermore, there appeared to be no major dif-ferences in aortic wall thickness or diameter betweenoim/oim and +/+ mouse aortas. In addition, there wereno obvious histologic abnormalities or differences in thelamellar or elastin organization between oim/oim and +/+mouse aortas.

DISCUSSION

We present evidence that the thoracic aorta of theoim/oim mouse exhibits decreased breaking strength andgreater extensibility in the absence of proα2(I) collagenchains and/or presence of the homotrimeric type I colla-gen isotype. The most pronounced biomechanical differ-ences were seen with circumferential extension of thedescending aorta. The average breaking strength (Fmax)circumferentially in the descending oim/oim mouse aortawas only 23% of the Fmax in +/+ mouse aorta. Similarly,the oim/oim mouse descending aorta had an IEM that was24% of the value in +/+ mouse aorta. The differences werenot as dramatic during circumferential extension of theascending aorta. Overall, the circumferential Fmax of theascending aorta (range, 132 to 219 g/mg) was muchlower than the Fmax of the descending aorta (range, 242-

Fig 5. Analyses of longitudinal maximum breaking strength (Fmax; a) and incremental elastic modulus(IEM; b) of descending aorta of wildtype (+/+; n = 7; black bar), heterozygous (oim/+; n = 3; diagonalbar), and homozygous (oim/oim; n = 5; open bar) mice. Values are expressed as mean ± SD in grams (Fmax)and grams per millimeter (IEM) per milligram of dried aortic tissue. *P ≤ .05; **P = .003.

a b

447 g/mg). The circumferential Fmax of the ascendingoim/oim aorta was approximately 60% of the value in the+/+ aorta. These findings are consistent with our under-standing that collagen and elastin bear the majority of thewall stress and determine the stiffness (compliance) of the aorta.2 With increasing distance of the vessel from theheart, the elastin content is known to decrease and the col-lagen content to increase, with a net result of higher collagen-to-elastin ratios.18,25

The aorta is an “elastic artery,” and subsequently itsmedia is composed of layers of smooth muscle cells inter-spersed with clearly defined lamellae of connective tissue.These lamellar units, composed predominantly of elastinand collagen lying in parallel, represent the functional andstructural unit of the aortic wall that responds to pressureand tangential stress and are responsible for the passivemechanical properties of the vessel.1,2 Several studies sug-gest that elastin determines the aortic compliance at lowphysiologic pressures and that collagen progressively takesover load-bearing at higher pressures.2,26-28 Accordingly,aortic stiffness also progressively increases from theascending and descending thoracic aorta to the abdominalaorta.18

In addition, the aortic wall is anisotropic, having dif-ferent mechanical properties in the three orthogonaldirections, circumferential, longitudinal, and radial.2Enzymatic degradation studies demonstrated that elastinbears the load in all three directions, whereas collagenbears the load almost exclusively in the circumferentialdirection, and smooth muscle cells do not significantlycontribute directly to the passive “static” biomechanicalproperties of the aortic wall.25,29-32 During the longitudi-nal stretching, all the aortic samples, regardless of mousegenotype, demonstrated much lower breaking strength(range, 13 to 42 g/mg) and higher extensibility (range, 6-

11 g/mm/mg) than when they were stretched circumfer-entially. These findings are also consistent with theanisotropic nature of the aortic wall. Moreover, the longi-tudinal extension showed no differences in the stiffness ofthe aortas in oim/oim and +/+ mice, which is also consis-tent with the decreased role of collagen along the longitu-dinal axis. However, to our surprise, there was still asignificant difference in Fmax longitudinally between thegenotypes, suggesting that type I collagen does influenceindirectly or directly the biomechanical properties alongthe longitudinal axis. One can easily postulate thatdeposits of the homotrimeric type I collagen isotypeand/or the absence of the proα2(I) collagen chain mayalter the fibrillar structure and architecture of the collagenfibrils and influence the structure and function of theother components of the lamellar units.

In addition, we found no significant differences in totalcollagen content or the morphological structure/organizationof the aortas of oim/oim and +/+ mice, suggesting that theabsolute amount of collagen is not the only factor that deter-mines arterial wall strength and further suggesting that theultrastructural collagen composition and organization of theaortic wall must play a significant role.

It is quite likely that all three components, elastin andtypes I and III collagen, contribute to the ultimate struc-tural integrity of the aorta. We postulate that a structuraldefect in any one or more of these components increasesthe load-bearing on the other components of the aorticwall and, thus, makes the vascular wall more susceptible tofatigue and failure. There is a great deal of clinical evidencefrom the pathogenesis of certain connective tissue disor-ders that supports this hypothesis. EDS types IV and VIresult from defects affecting type III and I collagens,respectively, and are strongly associated with decreasedarterial wall integrity and cardiovascular deficits, including

JOURNAL OF VASCULAR SURGERY1268 Vouyouka et al June 2001

Fig 6. Histology of media thoracic aortic sections of 7-month-old wildtype (+/+; A, B, C) and homozygous(oim/oim; D, E, F) mice stained with picrosirius red (A, D), hematoxylin and eosin (B, E), and Verhoeff-vanGieson (C, F; original magnification, 10×).

mitral valve prolapse, multiple arterial ruptures, and aorticdissections.19-21,33 Two unrelated individuals with classicalEDS (EDS type I/II) who have been previously describedsynthesize only homotrimeric type I collagen and are bio-chemically similar to the oim mouse22 and certain recessiveforms of human type III OI.15,16 Both individuals had car-diovascular complications, a grade IV aortic regurgitationand a severe mitral valve prolapse.12-14 However, no datawere presented about the integrity of their aortas. The sig-nificance of our findings with the oim mouse aortas in theclinical course of patients with aortic aneurysmal disease orwith type I collagen deficiencies is yet to be determined.The cardiovascular complications of OI, particularly inrelation to aortic diseases, have not been well character-ized. Only a handful of cases of spontaneous aortic dissec-tion directly related to OI are reported in theliterature.34-37 Vetter evaluated cardiovascular function in58 patients with OI types I, III, and IV and determinedthat patients with OI type III were more likely to haveectatic aortas.38 It is difficult to make any direct clinicalcorrelation with the oim/oim mouse, although it is clini-cally most similar to OI type III, because its moleculardefect is relatively rare (autosomal recessive inheritance ofa functionally null COL1A2 gene), with only a fewreported human cases.15,16

Deficiencies in elastin and collagen have also beenimplicated in the pathogenesis of aneurysmal degenerationin patients who do not have Marfan syndrome orEDS.2,17,39-41 Although there is a well-demonstrated cor-relation between elastin decrease and aneurysm formation,there still remains a lot of confusion regarding the role ofcollagen and the aneurysmal aorta in the literature; colla-gen concentrations have been found to be increased insome studies and decreased or unchanged in oth-ers.2,17,39,40,42,43

This study demonstrates the biomechanical conse-quences of exclusive expression of homotrimeric type Icollagen on the thoracic aorta of the oim/oim mouse andrepresents a unique model for investigating the role of theproα2(I) collagen chain in the structure and function ofthe vasculature and for defining the mechanisms and tis-sue-specific adaptations to over-expression of thehomotrimeric isotype of type I collagen. The oim mouseprovides a potential paradox for cardiovascular disease andaging. The increased extensibility of the oim/oim aortamay protect the oim/oim mouse against the developmentof elevated arterial pressures, but the oim/oim aorta withits significantly reduced breaking strength may also be sig-nificantly compromised when subjected to elevated arte-rial pressures or other vascular stressors. Long-termfollow-up of these mice during aging and under cardio-vascular stress, by using serial ultrasonic imaging of theiraortas, analysis of the aortic velocity profile under differ-ent conditions of systemic blood pressure, and analysis ofthe ultrastructure of the collagen should significantly con-tribute to our understanding of the role of the extracellu-lar matrix architecture in vascular function and disease.Finally, the greatest significance of the oim mouse model

JOURNAL OF VASCULAR SURGERYVolume 33, Number 6 Vouyouka et al 1269

is that it provides a foundation for future studies. In con-junction with other cardiovascular mouse models, such asthe elastin-deficient mouse,44 it will allow for very definedbreeding strategies, making it finally possible to systemat-ically evaluate the role of multigene interactions in vascu-lar and aneurysmal disease.

We thank Russell Stevens for his excellent technicalassistance in preparation of aortic samples.

REFERENCES1. Nichols WW, O’Rourke MF. Properties of the arterial wall. In:

McDonald’s blood flow in arteries. 3rd ed. Philadelphia: Lea andFebiger; 1990. p. 77-124.

2. Dobrin PB. Physiology and pathophysiology of blood vessels. In:Sidawy AN, Sumpio BE, DePalma RG, editors. The basic science ofvascular disease. New York: Futura Publishing; 1997. p. 69-105.

3. Morton LF, Barnes MJ. Collagen polymorphism in normal and dis-eased blood vessel wall: investigation of collagen types I, III, and V.Atherosclerosis 1982;42:41-51.

4. Byers PH. Disorders of collagen biosynthesis and structure. In: ScriverC, Beaudet A, Sly W, Valle D, editors. The metabolic and molecularbasis of inherited disease. New York: McGraw-Hill; 1995. p. 4029-77.

5. Phillips CL, Yeowell HN. Vitamin C, collagen synthesis and aging. In:Packer L, Fuchs J, editors. Vitamin C in health and disease. New York:Marcel Dekker; 1997. p. 205-30.

6. Moro L, Smith BD. Identification of collagen α1(I) trimer and nor-mal type I collagen in a polyoma virus-induced mouse tumor. ArchBiochem Biophys 1977;182:33-41.

7. Ghersi G, LaFiura AM, Minara S. Direct adhesion to type I andhomotrimeric collagens by breast carcinoma and embryonic epithelialcells in culture: a comparative study. Eur J Cell Biol 1989;50:279-84.

8. Rupard JH, Dimari SJ, Damjanov I, Haralson MA. Synthesis of typeI homotrimer collagen molecules by cultured human lung adenocar-cinoma cells. Am J Pathol 1988;133:316-26.

9. Jimenez SA, Bashey RI, Benditt M, Yankowski R. Identification ofcollagen α1(I) trimer in embryonic chick tendons and calvaria.Biochem Biophys Res Comm 1977;78:1354-61.

10. Uitto J. Collagen polymorphism: isolation and partial characterizationof α1(I)-trimer molecules in normal human skin. Arch BiochemBiophys 1979;192:371-9.

11. Kay EP. Rabbit corneal endothelial cells modulated by polymor-phonuclear leukocytes are fibroblasts. Invest Ophthalmol Vis Sci1986;27:891-7.

12. Sasaki T, Arai K, Ono M, Yamaguchi T, Furuta S, Nagai Y. EhlersDanlos syndrome: a variant characterized by the deficiency ofproα2(I) chain of type I procollagen. Arch Dermatol 1987;123:76-9.

13. Kojima T, Shinkai H, Fujita M, Morita E, Okamoto S. Case reportand study of collagen metabolism in Ehlers Danlos syndrome type II.J Dermatol 1988;15:155-60.

14. Hata R, Kurata S, Shinkai H. Existence of malfunctioning proα2(I)collagen genes in a patient with a proa2(I) chain defective variant ofEhlers-Danlos syndrome. Eur J Biochem 1988;174:231-7.

15. Nicholls AC, Osse G, Schloon HG, Lenard HG, Deak S, Myers JC,Prockop DJ, et al. The clinical features of homozygous α2(I) colla-gen-deficient osteogenesis imperfecta. J Med Genet 1984;21:257-62.

16. Pihlajaniemi T, Dickson LA, Pope FM, Korhonen VR, Nicholls A,Prockop DJ, et al. Osteogenesis imperfecta: cloning of a pro-α2(I)collagen gene with a frameshift mutation. J Biol Chem 1984;259:12941-4.

17. Sumner DS, Hokanson DE, Strandness DE Jr. Stress-strain character-istics and collagen-elastin content of abdominal aortic aneurysms.Surg Gynecol Obstet 1970;130:459-66.

18. Clark JM, Glagov S. Transmural organization of the arterial wall; thelamellar unit revisited. Arteriosclerosis 1985;5:19-34.

19. Pope MF, Burrows NP. Ehlers-Danlos syndrome has varied molecularmechanisms. J Med Genet 1997;34:400-10.

20. Beighton P, De Paepe A, Steinmann B, Tsipouras P, Wenstrup RJ.

Ehlers-Danlos syndromes: revised nosology. Am J Med Genet 1998;77:31-7.

21. Byers PH. Ehlers-Danlos syndrome. In: Rimoin DL, Connor JM,Pyeritz RE, editors. Emery and Rimoin’s principles and practice ofmedical genetics. 3rd ed. New York: Churchill Livingston; 1996. p.1067-81.

22. Chipman SD, Sweet HO, McBride DJ, Davisson MT, Marks SC,Shuldiner AR, et al. Defective proα2(I)-collagen synthesis in a reces-sive mutation in mice: a model of human osteogenesis imperfecta.Proc Natl Acad Sci U S A 1993;90:1701-5.

23. Phillips CL, Bradley DA, Schlotzhauer CL, Bergfeld M, Libreros-Minotta C, Gawenis LR, et al. Oim mice exhibit altered femur andincisor mineral composition and decreased bone mineral density. Bone2000;27:219-26.

24. Stegemann H, Stadler K. Determination of hydroxyproline. ClinChim Acta 1967;18:267-73.

25. Dobrin PB, Baker WH, Gley WC. Elastoytic and collagenolytic stud-ies of arteries: implications for the mechanical properties ofaneurysms. Arch Surg 1984;119:405-9.

26. Matsumura S, Kawazoye S, Tian SF, Takashima T, Sunaga T, FujitaniN, et al. Organizations of extracellular matrices in aortic and mesen-teric arteries or stroke-prone spontaneously hypertensive rat. Ann N YAcad Sci 1995;748:534-7.

27. Greene MA, Friedlander R, Boltox AJ. Distensibility of arteries inhuman hypertension. Proc Soc Exp Biol Med 1966;121:580-5.

28. Krafka J. Changes in the elasticity of the aorta with age. AMA ArchPathol 1940;29:303-9.

29. Dobrin PB, Canfield TR. Elastase, collagenase and the radial elasticproperties of dog carotid artery. Am J Physiol 1984;247:H123-31.

30. Bank AJ, Wang H, Holte JE, Mullen K, Shammas, Kubo SH.Contribution of collagen, elastin, smooth muscle to in vivo humanbrachial artery wall stress and elastic modulus. Circulation 1996;94:3263-70.

31. MacLean NF, Dudek NL, Roach MR. The role of radial elastic prop-erties in the development of aortic dissections. J Vasc Surg 1999;29:703-10.

32. Roach MR, Burton AC. The reason for the shape of the distensibilitycurves of arteries. Canadian Journal of Biochemistry and Physiology1957;35:681-90.

33. Wenstrup RJ, Murad S, Pinnell SR. Ehlers-Danlos syndrome type VI:clinical manifestations of lysyl hydroxylase deficiency. J Pediatr1989;115:405-9.

34. Isotalo PH, Guindi MM, Bedard P, Brais MP, Veinot JP. Aortic dis-section: a rare complication of osteogenesis imperfecta. Can J Cardiol1999;15:1139-42.

35. Ashraf SS, Shaukat N, Masood M, Lyons TJ, Keenan DJ. Type I aor-tic dissection in a patient with osteogenesis imperfecta. Eur JCardiothorac Surg 1993;7:665-6.

36. Moriyama Y, Nishida T, Toyohira H. Acute dissection in a patientwith osteogenesis imperfecta. Ann Thorac Surg 1995;60:1397-9.

37. Cusimano RJ. Repeat cardiac operation in a patient with osteogenesisimperfecta. Ann Thorac Surg 1996;61:1294.

38. Vetter U, Maierhofer B, Muller M, Lang D, Teller WM, Brenner R,et al. Osteogenesis imperfecta in childhood: cardiac and renal mani-festations. Eur J Pediatr 1989;149:184-7.

39. Powell J, Greenlalgh RM. Cellular, enzymatic, and genetic factors inthe pathogenesis of abdominal, aortic aneurysms. J Vasc Surg 1989;9:297-304.

40. Menashi S, Campa JS, Greenhalgh RM, Powell JT. Collagen inabdominal aortic aneurysm: typing, content, and degradation. J VascSurg 1987;6:578-82.

41. Baxter BT, MaGee GS, Shively VP, Drummond IA, Dixit SN,Yamauchi M, et al. Elastin content, cross-links, and mRNA in normaland aneurysmal human aorta. J Vasc Surg 1992;16:192-200.

42. Rizzo RJ, McCarthy WJ, Dixit SN, Lilly MP, Shiverly VP, Flinn WR,et al. Collagen types and matrix protein content in human abdominalaortic aneurysms. J Vasc Surg 1989;10:365-73.

43. Thornell LE, Norrgard O, Eriksson A, Vanderwee M, Angqvist KA.Abdominal aortic aneurysms: distribution of elastin, collagen I andIII, and intermediate filament proteins desmin and vimentin-a com-parison of familial and nonfamilial aneurysms. Heart Vessels 1986;2:172-83.

44. Li DY, Brooke B, Davis EC, Mecham RP, Sorensen LK, Boak BB, etal. Elastin is an essential determinant of arterial morphogenesis.Nature 1998;21:276-80.

Submitted Jul 4, 2000; accepted Dec 5, 2000.

JOURNAL OF VASCULAR SURGERY1270 Vouyouka et al June 2001